Comment by rstuart4133 | Hacker News Reader

Comment by rstuart4133 | Hacker News Reader

Parent Full thread

> Problem with 3D printing is how slow it really is.

Yeh, it's slow. But the article is talking about "The ability to build a one-off product such as a boat or small yacht, directly from CAD".

I dunno about you, but I've designed and paid for injection molds, and there is nothing cheap or fast about producing the first one. We were a small company, those molds cost us a substantial fraction of our capital. I would not describe the feelings engendered by unpacking and checking that first one as magical. Words that spring to mind are hope and dread, and eventually relief when we realised it would not be that expensive to move through holes in PCB's by 0.5mm, and adjust metal jigs, and resignation that the aesthetics of the thing were not quite as we expected, but would do.

Now a technology that allowed us to print something in 24 or 48 hours for a few thousand that was actually useable in the field - that would have looked absolutely magical. Being able to print something size of a boat hull that is strong enough to sail is almost unbelievable, even now.

More realistically you would one-off something like a boat out of fiberglass or carbon fiber and a non-metal mold (either shaped wood or foam, both of which can be large-scale milled easily). Yes, it would be more expensive than 3D printing, but a hell of a lot cheaper than an injection mold and with far better material properties than any thermoplastic. Or do it with vacuum forming. Injection molds are expensive because they are specifically designed and made for high-volume production with tight tolerances. If you don't need those things there are much cheaper options.

I'm personally bullish on 3D printing long-term, but FDM mostly still sucks and gets a lot more positive press than actual practical use in industry. It's just the easiest and cheapest of the printing technologies to expand on, so you see a lot of concept projects that sound good on paper.

The more interesting comparison point might be other CNC machining techniques which share the ability to create one off products directly from CAD with raw material, but seem to get a fraction of the attention of 'printing'. Leaving aside them being less suited to some use cases than 'printing' (and much more suited to others), is computer controlled manufacturing using an additive process really that much more magical than computer controlled manufacturing using a subtractive process?

You'd be throwing an awful lot of material away if you tried to machine a boat out of a solid block of plastic.

Subtractive manufacturing can give you a higher-quality result and can use a wider range of materials, but the machinery costs more because it needs to be stronger, and the materials cost more because you throw away everything that doesn't go into the final part.

Surely if you do it at scale the removed material can be recycled fairly cheaply?

Standard practice in CNC machining (at a certain scale) to deliver at least aluminum back to the distributor. Not all materials can be recycled as effectively though, or have a price point where it is worth doing. Plastic has the problem of being incredibly cheap.

Is there anyone with a metallurgy background reading this? It's easy to melt aluminium at home, what are the properties of a billet cast with primitive tools vs in a factory? Like, is it more brittle, or does it have cavities etc? What circumstances does a factory control better for to get a better product?

For garage level, I built this charcoal aluminum forge with my friend and it works great for the price! https://www.youtube.com/watch?v=hHD10DjxM1g

It's relatively easy to re-cast aluminum at home at small scale. I think the biggest challenge for boat-sized would be the massive amount of energy needed to heat all that metal, plus the huge crucible and machinery to pour it.

Another issue with home-cast aluminum is alumina (oxidized aluminum) which will make small inclusion defects in your print. I've had several casts fail because I wasn't able to properly separate the alumina and the high-alumina areas were very brittle. I think most aluminum foundries have the electrolysis capabilities to convert all the alumina back to aluminum, but this is an extremely high voltage hot process that does not seem feasible for garage-scale.

(As an aside, aluminum cans have a ton of oxide because of the high surface area, so you get a ton of dross that must be skimmed off to use the molten aluminum. We would do it in two steps: first melt cans and pour in a muffin tin to make small ingots, then use those higher-purity ingots to cast the desired piece. Probably better in most circumstances to just start with lower-surface-area aluminum materials: better to pay for quality bar stock or ingots rather than spend the same on more charcoal for multiple runs.)

Note: if anyone tries this, invest in safety! At minimum, you should have welding gloves, sturdy tongs and a face shield. You should understand what happens when molten aluminum comes in contact with moisture (wet concrete is extremely dangerous!), practice your pouring motion with a cold crucible, and keep a bucket of water and/or fire extinguisher nearby in case you start a fire.

But please research casting safety yourself, I'm no expert. 1200+F temperatures significantly more dangerous than the high temperatures we interact with in our daily lives.

Casting is a massive pain in the ass. It's one of those cases where the difference is in all the sum of all the little things and refinements that only make sense at scale.

Even at home/hobbyist scale, it's worth taking aluminium swarf to a metal recycler.

Plastic scrap value is near zero [1] when compared to Aluminum

[1] https://webcache.googleusercontent.com/search?q=cache:fZ0f_O...

Plastic of unknown provenance is worth nearly nothing.

Clean pure plastic chips that just got cut off of a block of new plastic can be made back into pellets and used for new stuff and is worth a little less than plastic pellets.

I could see this working at scale, but nobody is going to trust a hobbiest or small shop that they haven't contaminated the bin.

Indeed, I work for a company that manufactures high end technology, and we're in a comfortable position where we don't need the production volume of injection molding, and can use soft-tooling processes. It's a lot less nerve wracking. Also, paint hides a lot of sins. ;-)

The thing to be careful about: Don't assume you know what's cheaper: Injection molding or something else. I've had my own assumptions blown away so many times.

But 3d printing is still kind of at the phase that machine learning went through, where every meeting inspires at least one manager to blurt out: "Have you considered using 3d printing for this design?" with the other managers secretly wishing that they had been so clever.

I think there are different use cases, from an fully automated Lego factory to a fully automated clay-house printer on Mars.

Reply on news.ycombinator.com